Delivery of chromosomes, the basic units of inheritance, to each daughter cell during cell division is mediated by the centromere. Unlike typical genes, in metazoans this central genetic element is not determined by DMA sequence. Rather, functional centromeres are determined epigenetically through stable acquisition of an unexplained, non-DNA "mark". A prime candidate for a component of such an epigenetic mark is CENP-A, a histone H3 variant found exclusively at functional centromeres. We will determine how incorporation of CENP-A affects the underlying structure of the nuclesomes into which it assembles. This will be done using deuterium exchange mass spectrometry to identify the structural characteristics and conformational rigidity of nucleosomes assembled from 1) CENP-A, 2) histone H3 or 3) histone H3 carrying the centromere targeting domain of CENP-A. To identify how CENP-A may function in centromere assembly, components directly recruited by CENP-A, and potential cell cycle-dependent modifications of CENP-A, will be identified by mass spectrometry after purification of centromeric nucleosomes assembled in vivo with affinity tagged CENP-A. Similar affinity-tagged CENP-A nucleosomes assembled in vitro may also be used to purify components that selectively target to such nucleosomes. The timing and dynamics of centromere assembly before and after replication will be determined using fluorescently tagged CENP-A variants and photobleaching methods (FRAP and FLIP) or with tetra-cys CENP-A and fluorescent arsenic compounds (FLASH). Methods will be developed for rapidly following de novo centromere assembly in mammalian cells after introduction of large arrays of centromeric alphoid DNA carried on bacterial or yeast artificial chromosomes. These will be used to identify factors critical for formation of new centromeres. The medical implications of understanding how centromeres function and the genetic mechanisms that may generate failure of normal chromosome delivery are broad. Among these, errors of chromosome segregation lead to infertility. Moreover, many human tumors have highly abnormal numbers of chromosomes (that is, they are aneuploid), with initial chromosomal loss participating in the early steps of the transformation cascade in inherited cancers caused by heterozygous mutation in tumor suppressor genes and the more widespread aneuploidy characteristic of advance tumors thought to drive acquisition of malignant growth properties. [unreadable] [unreadable]